E50: The Future of Crypto with Diogo Mónica of Haun Ventures

0:05Welcome back to Turpentine VC, a podcast where we discuss the art and science of building successful venture firms, VC to VC. Today, we are joined by Diogo Monica. Diogo is a general partner at Han Ventures, a crypto venture firm founded in 2021 by former A16Z general partner Katie Han, who co-led Andreessen's first three crypto funds. Diogo and I discuss the evolution, current state, and future of the crypto ecosystem, especially as it innovates infrastructure for more traditional industries. We examine the intersection of AI and crypto, the impact of blockchain technology on fintech, and the future of crypto venture.

0:42Please enjoy. Diogo, welcome to Turpentine VC. Thanks for coming on the podcast. Thank you so much for having me, Eric. Diogo, why don't you give background of what Han Ventures is and how you came to join it? Yeah, Han Ventures is crypto funds. It was started two years ago by Kitty Han. Kitty is an amazing human being, but also has been very known in the crypto industry. I think, as you know, she was a former federal prosecutor. She was on the board of Coinbase. She was on a series of other tech boards. She ended up co-leading A6 and Z Crypto Fund 1, 2, and 3 before striking out on her own and starting a new fund, aptly named Han Ventures.

1:25It's actually two funds, not just one, but it's a$500 million early stage fund and a billion dollar early stage fund. And we invest, as I mentioned, primarily in crypto companies and digital asset companies. And that's awesome. And so here's a prompt I want to do with you. Imagine that I went in a coma in 2017 and I woke up in 2024. And in 2017, I was pretty up to date with what was happening in the world of crypto as it relates to VC at the time. people were saying, hey, is crypto going to change how VC operates on a sort of fundamental level, on a structural level? But then also just how big is this asset class going to be?

2:03Where are the opportunities going to be? Who's going to win? Is it specialized firms? Is it generalized firms? Why don't you give a little bit of an overview to my coma self who just woke up after seven years? Yeah. So as far as the size of the asset class, I will quote Matt Kohler and say that My job is not to predict the future, but to see the present clearly. And, you know, no comments on that one. I do think larger than the present, but nobody knows what the terminal size of this one. Yeah, so the terminal size. But in general, in terms of what has happened in the past few years, I do think the main thing that has happened is, as you know, 2017 actually barely, we barely had Ethereum.

2:42We had lots of things that were happening, but really we were going and starting this phase of smart contract blockchains instead of just Bitcoin as, you know, store of value. Let's call it this way. Lots of narratives that can be put on Bitcoin, the store of value narrative. And so that smart contract platform really can do anything. Smart contracts are programs, programs are Turing complete. And so they can definitionally do everything that you can do. However, this is no fault of Ethereum. It's just how it started. there was absolutely no scaling ability. There was a smart contract programming language that was a pretty unfortunate early choice that led to lots of hacks and lots of bugs.

3:19There was lack of infrastructure, there was lack of, in certain cases, liquidity for decentralized Xs or these other types of decentralized exchanges. But you can compose and create new things on top. And so now, this many years later, all of those things are done. We have and continue exploring, candidly, like all of the weird nooks and corners of distributed systems. And we have platforms that can do hundreds of thousands of transactions per second. We have lots of different clear use cases and integrations that allow you to go in and out of crypto. We have way more sophisticated on-ramps, so it's a lot easier than ever to turn dollars into crypto and crypto into dollars, despite the fact that the regulators have been extremely good at pushing back on banking crypto companies.

4:03With that asterisk aside, there's a lot more that has been built. There's a lot more that you can build on top of. And I know, you know, Dan Romero, for example, I think that's a good example of something that took a while to mature. It's built on base. Base didn't exist. L2s didn't exist. Optimism didn't exist. All of that is a sequence of events that had to happen for something like Farcaster to actually be able to be built. And flesh that out a little bit. Why did they have to exist for something like Farcaster to be built? So specifically, not specifically for Farcaster, but in general, I think the number of transactions that you have per second, the popularity being the curse of Ethereum.

4:36The more popular your program is or your smart contract is, the higher the fees go. If you have a blockchain that has tens of transactions per second, then definitionally, every time you're trying to settle to the actual blockchain itself, you're going to have contention. You're going to have rising fees. And so you can't really do the use cases that you want it to do. And right now you can't. There's a lot of competing platforms, but also Ethereum itself is a very sophisticated distributed system that is open source that somehow has been able to evolve. You've had massive evolution, not just from proof of work.

5:06So using mining as the way that you get to consensus, but to proof of stake. That's an extremely complex piece of engineering that actually was flawlessly executed and deployed. And lots of different innovations are happening for you to have all of these different aspects. And in particular, scaling as number of transactions per second, the inability of people of accessing the ecosystem, easy, either developers or users. So that means more wallets with greater UX, with ease of onboarding, where you can kind of have companies like Privy, for example, that just abstract away the complexity of generating wallets.

5:42And on the developer side, SDKs and the ability to really be building new applications, analytics companies. All of these is just, they're all there. They weren't there. In fact, we're potentially through the second wave of this infrastructure company. and we've gone in so many different directions that you can build upon, so many different blockchains, so many different layer twos, so many different ways of you having data availability, so many different ways of you having borrowed security from different blockchains that right now it's actually like, it has to come back. So it has to come back to something that is more sane in terms of optionality because right now developers do have lots of options on how to build these Web3 companies and these protocols on top of each other.

6:22Yeah, let's zoom out. Remember, I just woke up from 2017. Can you sort of make like a bit of a high level market map in terms of like how you bucket the different types of investable opportunities in the asset class? So, yeah, if we zoom out in terms of buckets, I think one of the things that you'd notice that you didn't have in 2017 is, of course, the institutionalization of Bitcoin. What I mean by that is Bitcoin ETFs are now available to any American citizen that has a brokerage account. And so you can just search for BTC and you have one of the world's, if not the world's largest financial company, BlackRock, offering an ETF.

7:00That is a massive, massive change. The institutions have gone through all the hurdles that they needed to go through to actually be able to play in the ecosystem. And it's not just a Bitcoin. Bitcoin people sort of like, if in 2017, I told you that there was going to be a Bitcoin ETF, the majority of people wouldn't believe you. But if they were to believe that there was an ETF, then it would be Bitcoin. But here we are. There's going to be an Ethereum ETF very soon. And then if you kind of like map that out, the opening of the regulatory space and the opening of the institutionalization of crypto has happened in 2024 right now that you just could not make a claim that was happening really in 2017.

7:36It was all about institutions are coming and you can hear the stampede. But, you know, it took us up until 2024, really, for that stampede to actually come to fruition. And it was really a stampede. It was a little bit more of a trickle that then one day you wake up and it's like a boiling frog or boiling water. like, oh, actually, they're all here now and all of them have a play. So that is one bucket, which is kind of like this institutionalization of crypto with ETFs. Within the same bucket, I would say that there's narratives that have been with us from day one of crypto, things like real world assets.

8:08We have wanted to tokenize everything since day one. And for some reason, it keeps not happening. And right now you actually do have BlackRock tokenizing treasuries and you have Biddle in this case, which is a way better way of getting exposure to these money market funds because you have 24-7 instant liquidity on stable coins. You can move around. So if on a Friday night you want to get liquidity on your T-bill, you don't have to wait for Tuesday morning, right? So you have to wait the whole weekend and then T plus two to actually, or T plus one at a minimum to get your funds. So in that marketing category, we have all the institutions are playing.

8:45They have specific plays. They have issued on public chains, there are RWAs and real RWAs. And yes, it's just the beginning. Yes, ETFs really that are out are just the Bitcoin. But you have a very legible path for A, more ETFs, B, more assets tokenized in RWAs on chain. So that's a really good one. Another one would be other elements of product market fit. In 2017, we were very early in the stablecoin game. Right now, you have stablecoins as probably the killer use case for crypto right now. and with very clear product market fit. In 2024, you woke up and there's over$160 billion of stablecoins issued.

9:25In fact, stablecoins collectively are one of the largest holders of US treasuries. I believe they're actually the 18th largest holder of US treasuries. They're actually larger than Germany. They hold more US treasuries than the whole country of Germany, which is kind of an insane stat. And that is not slowing down in terms of growth. and it will continue growing. So yet another bucket of programmable money is something that people, developers, and just everywhere in the world, people are demanding. It has pretty cool and crazy use cases that people are doing it with. One of which is just candidly, just going back to our FinTech days and all of the hype of FinTech, but just rebuilding it without having to wait 18 months or two years for authorization for you to actually have access to financial rails.

10:10And that's actually game changer for entrepreneurs, for people building companies and incredibly exciting for a venture investor, of course. Then we can also go on what has happened in DeFi and how that has evolved, what has happened in, honestly, architecture of blockchains and how layer twos and zero knowledge have actually allowed you to scale a lot of this core underlying infrastructure in a very successful way. We can also talk to you about what is different between NFTs today and NFTs then and how the hype kind of came and went. There's lots of different buckets that we can like slot on, but I would say the three biggest ones are the infrastructure piece, are the DeFi on how the composability has happened and what has actually been successful there, the stable coins as programmable money, allowing anyone to build any kind of company on top of US dollars, and obviously the institutionalization with RWAs and with ETFs being the two biggest ones.

11:02What do you think is something that feels underrated in terms of maybe something we haven't yet spoke about that mainstream isn't talking about? So I feel like and I've always felt that there's a big element of just positive externalities of crypto that we don't really talk about that much. And, you know, part of the reason for it is because we're all it's all lost in the meme coin mania and the ICO boom and whatever is the topic du jour. But I do think a lot about these positive externalities of crypto. So in one way, one way to look at this is that the crypto VC community has actually poured billions of dollars into deep cryptographic research and zero knowledge, fully homomorphic encryption.

11:49We've even somehow made Reed Solomon codes cooling it, which, you know, there's a rise of the erasure coding for decentralized file storage. So it's kind of fascinating to think through that we have poured so much money into fundamental cryptographic research and how the reality of it is that the majority of it won't benefit just crypto. It is really about everything else that can be built with it. All of these zero knowledge and fully homomorphic encryption are going to have more impact in just normal corporates, database technology, which affects everyone, really than they're going to have in crypto, which is a subset of its application.

12:26So that's kind of interesting for you to think through of we are funding fundamental research that is going to be used outside of crypto. And so as a venture investor, if the research is happening here, people are using it, of course, for what's right in front of them. So they're inventing fully homomorphic encryption or zero knowledge proofs, but then they're using it to do some settlements with Ethereum layer two technology back onto Ethereum without realizing that, wait a minute, this particular piece of technology, Maybe it's a lot better for an identity scheme that is actually private or actually, you know, a CBDC that allows you to have the same types of components of cash with the ability of the banks maintaining their actual compliance elements on top of while still maintaining the privacy aspects that the cash has and without having to carry money or really around just general identity.

13:16There are so many places on the internet that we have to share identity with that just didn't have to be. And those are not crypto use cases at all. They're just normal, traditional run-of-the-mill interactions that we have on the internet on a daily basis that could make MADE a lot better. So I do think a lot about these ones. Another one that is obviously a lot on our minds are these use cases around misinformation and the appearance of perfectly real JNAI videos that just obviously are completely fake or completely out of context. and how we just have the perfect technology to actually help with this, help with origin.

13:57And even cameras these days are now coming with private keys that allow you to actually sign pictures and potentially some of the videos that are coming out of them. And we created a technology that is amazing at coordination of multiple distressing parties. And it would be fantastic that we would actually use it to solve this problem that is actually appearing. And so all of these things of externalities of crypto and actually crypto technologies play outside of crypto. I don't think a lot of people look to because, you know, you just use them for whatever in crypto gives you the most money, which usually is launching a token or launching an effort.

14:31And so that's usually what it gets used for. Hey, we'll continue our interview in a moment after a word from our sponsors. How deep do you go to seek out an answer to a question? Maybe you've spent hours clicking the source links on an obscure Wikipedia page, or maybe you're even the type of person who checked out the entire shelf on the topic at your library. If you're nodding along, then check out GiveWell, an organization that researches questions about global health and philanthropy. Even if a satisfying answer might require years of reviewing studies, talking to experts, and over 300 footnotes.

15:00GiveWell has now spent over 17 years researching charitable organizations and only directs funding to a few of the highest impact opportunities they've found. Over 125 ,000 donors have used GiveWell to donate more than$2 billion. dollars. Rigorous evidence suggests that these donations will save over 200 ,000 lives and improve the lives of millions more. GiveWell wants as many donors as possible to make informed decisions about high-impact giving. You can find all of their research and recommendations on their site for free. You can make tax-deductible donations to their recommended funds or charities, and GiveWell doesn't take a cut.

15:33If you've never used GiveWell to donate, you can have your donation matched up to$100 before the end of the year, or as long as matching funds last. To claim your match, Go to givewell.org and pick podcast and enter econ102 with Noah Smith and Eric Torenberg at checkout. Make sure they know that you heard about GiveWell from econ102 with Noah Smith and Eric Torenberg to get your donation matched. Again, that's givewell.org to donate or find out more. What do you think are the things that haven't panned out yet that me in 2017 might have thought, hey, ideally we would have had this by now. Or that in 2034, or that is a long time from now, but that you think, hey, this is what the real promise is.

16:17What hasn't yet panned out that you think has massive opportunity that maybe ideally would have by now? Actually, in every single vertical, things haven't panned out exactly as we wanted them. That's just the case. I think maybe outside of sovereign resistant money, which has panned out exactly how people hoped, maybe even more than they hoped to get, and maybe outside of stablecoins. So let's talk about RWAs again. RWAs right now are very constrained towards money markets. But really, the whole goal was about tokenizing equities, tokenizing bonds, tokenizing fixed income and having the ability of having smart contracts manage these baskets and do new financial instruments and having unfettered access to these financial instruments without actually having to go through the traditional hurdles that the financial companies make you go through.

17:07That really hasn't happened. We are in the early beginnings of it and we do see the path, but it's still pretty far off. And so tokenizing money markets is very far off from tokenizing an index fund and having it available to an American citizen in the United States of America. Right now, that future seems a little bit far off. And even in our daily ways, it's mostly about institutions getting more liquidity or faster access to their capital. That's the main use case, which is really not what we've been promising everyone for a long time. Another one that has just been, you know, pet peeve in mind is the whole putting real estate on the blockchain or like loans on the blockchain, mortgages on the blockchain.

17:43These types of use cases that just require so much paper pushing and real world interaction. So many in the United States alone, every single state has its own little database of where things come from. They're just so complex and so hard to do without the ability of actually scaling worldwide. And those have kind of been left by the wayside. And the majority, if not all, of those attempts have practically failed. And it's still yet something that people are desperate for. And so the pain is very high, but somehow the solution and the specific products that have tried to actually ameliorate this have really been able to crack it, crack that nut.

18:16So there's lots of different things like that that I think should be pointed to as we're not there yet. And 10 years ago, we said we would be because 10 years is a long time. And what about this sort of, you know, there was this dream of, you know, you mentioned Forecaster, but decentralized social networks, decentralized marketplaces, you know, that there were, you know, Uber, Airbnb, Facebook, that the next iterations of kind of the big Web 2 companies would be decentralized. What have we learned about, you know, how has that panned out, which are likely to be Web 3 and which are likely to not be?

18:54So in general, something that has been very obvious is that people don't particularly care about privacy and they're definitely not willing to go through hoops to get it. And so that's just one of the learnings that many people that have tried to go towards these use cases have found out. The other thing is that beating Uber means that you have to be much better than them. And you can't do that if you're already starting with both your hands tied behind your back with some technological disadvantage, which in many cases, decentralizing a platform by definition is going to make it less efficient than a centralized one because you just have a different set of different set of goals.

19:31And your goal is the ability of, I don't know, having a Byzantine full tolerance and tolerate three F plus one in which F is malicious attackers. And Uber does not have that. They trust their systems completely. And so if they have less restrictions, they move faster. So it's as simple as that. And so in a lot of cases, the me too types of products were always doomed to fail. Think about it this way. If I tell you that I want to create a new Uber and I tell that I'm going to do it with a technology that is less proven, harder to scale, slower, has all these advantages. And the only thing that I'm going to have as an answer as to why it's going to be better is because there's alignment of incentives and I'm going to be able to overcome the cold start problem.

20:10I would say, yes, maybe you can overcome the cold start problem and get some early excitement. And tokens are amazing at that. But you won't be able to maintain it. You will have churn. It's incredibly hard to run this business. It is a massively operations-focused business. And you need to be a world-class founder with a world-class team to solve it. So nobody thought that just because you had crypto into it could just solve these problems. And that brings me to actually, we talk about tokens. Tokens are amazing. They're amazing in many ways. and they have a superpower, but they also have this super weakness.

20:41The superpower really is that you can overcome the cold start problem that is hard for many startups, which is you get these network effects kicking. You have an artificial way of doing so through alignment of incentives, through token issuance. You can actually get people behaving in the ways that you want it to. That's amazing. However, they have a super weakness, which is, especially for a venture investor, they hide away per market fit. You can't really see per market fit through the mess of what's happening. And when you're looking at it and numbers go up, Then all of a sudden you realize that there's some fundamental thing that's broken.

21:11There's no retention. And traditional metrics sometimes help you, sometimes don't help you. And crypto metrics definitely don't help you in many cases. TVL and things like that are very gameable. It's very hard to judge daily active users when it's on chain and it's very easily fakeable. And all these addresses kind of look the same because you can generate as many of them as you want. So all of a sudden you have the super weakness of the product market fit is not obvious while having the superpower. And so the use cases for which that work, we've seen some of them, but it's not going to be magical for you to be able to compete against these behemoths of companies that have had amazing products for over a decade now and enrich and sprinkle some tokens into it and somehow win.

21:51Talk about AI and how you think AI has or will intersect with crypto. Look, AI and crypto, very candidly, the main thing working in AI and crypto is a narrative. It makes sense. People are excited. I get it. It's kind of perfect, actually. If you think about it, AI creates digital abundance and crypto enforces digital scarcity. So they're kind of like made for one another. But if you're seeing the present clearly, you realize that there's not a lot of it that is truly working. Maybe none of it is truly working yet. And so if you think about it, look, we first talk about decentralized training.

22:32It doesn't really make sense to have decentralized training in the sense that you're actually decentralizing the training of the models themselves, for example. And definitely not in a world where access to hardware is actually one of the scarce goods. So it's scarce. You can't access it. So it becomes very hard for you to create your decentralized system in the first place because you're competing with everybody else for the same types of resources. And then the other side is that maybe it's not decentralized training. It's actually a marketplace that uses crypto for decentralized training. And in which case, that's not really necessarily what I would call decentralized training, right?

23:06That is using crypto incentives and maybe just stable coins to facilitate a marketplace in which your GPUs are accessible. And so, yeah, it uses crypto to a certain extent, but it's not really a crypto use case. So I don't think those are necessarily fair. Maybe they go into crypto because they add a token and they force you to pay with the token and now the token has utility. So if you have a token, by definition, you're crypto. But really what they are is an attempt at a marketplace where you're adding GPUs and people are going to use these GPUs. And so that's not, not, not particularly interesting as a use case itself.

23:33And then we can think about decentralized inference and at the face of it, it makes more sense, right? There are LLMs that you might run that have been trained on data and are generating outputs that it would be quickly taken down if it was in a centralized fashion. Let's just put it that way, right? You train these models on data and just producing outputs that is very close to something that was, so you're clearly going to get a tick down. And so you enter this like gray area of, should I be able to run a model that is trained on, I don't know, Disney data locally on my laptop and do whatever I put I want.

24:08Clearly I'm not going to be able to host it on Hugging Face, right? And so that's an interesting one in which there's other gray area. There's not as clear, like breaking the law of that, that would be like, sort of like BitTorrent had Linux as distribution as the, as the, the good use case. I'm sure there's a good, very clean use case for it. But all the other ones that require decentralized inference are use cases in which you have to make some more assumptions about what's actually being what these models have been trained on. And then the other places where you need trust, there's definitely a lot to be said about, can you trust the outcome of the models?

24:43And can you use cryptographic mechanisms to guarantee that these outputs are correct? So there's lots of people trying to use in-zero knowledge, being able to prove that the models are correct. But look, UADs, three, four orders magnitude of overhead in every single one of these things. And I never want to bet against engineers, so we'll only get better. But for now, I think just traditional incentives around if the model has a bad output, the client will stop using it and paying for it. So it's a lot simpler for you to sort of like just the ranking system to judge a lot of these things. And I'm sure there's a lot of like details and elements in which the model output being provable and being cryptographically sound is absolutely required, but no use cases like that really exist right now.